Materials Transactions, Vol. 44, No. 4 (2003) pp. 504 to 510
Special Issue on Platform Science and Technology for Advanced Magnesium Alloys, II
#2003 The Japan Institute of Metals
High-purity Magnesium Coating on Magnesium Alloys
by Vapor Deposition Technique for Improving Corrosion Resistance
Harushige Tsubakino1 , Atsushi Yamamoto1 , Shinji Fukumoto1 , Atsushi Watanabe1; *1 ,
Kana Sugahara1; *2 and Hiroyuki Inoue2
1
2
Graduate School of Engineering, Himeji Institute of Technology, Himeji 671-2201, Japan
Graduate School, Osaka Prefecture University, Sakai 599-8531, Japan
Microstructures in coated magnesium alloy with high purity magnesium fabricated by applying a vacuum deposition technique were
investigated. Moreover, relationships between microstructures in coated and un-coated magnesium alloys and corrosion behaviors were
interpreted by in-situ laser microscopic observations during salt immersion tests. Magnesium with 3N-grade and AZ31 magnesium alloy were
used for an evaporation source and a substrate for deposition. Temperature of the substrates was changed resulting in change in temperature
profile in a furnace in order to optimize deposition coating conditions for obtaining homogeneous microstructures and thickness in deposited
layer. The coated specimen revealed superior corrosion resistance to those on 3N–Mg, AZ31 and AZ91E alloys, and comparable to that on 6N–
Mg in salt immersion tests using 3% NaCl solution at 300 K for 587 ks. In-situ observations showed that inhomogeneity in microstructures, such
as second phases and grain boundary segregations, deteriorate corrosion resistance in magnesium alloys. Therefore, pure magnesium coated
layer without inhomogeneity in metallographic and electrochemical meanings can improve the corrosion resistance on magnesium alloys.
(Received October 23, 2002; Accepted January 20, 2003)
Keywords: magnesium alloy, corrosion rate, deposition, coating, in-situ observation, corrosion behavior
1.
Introduction
Poor corrosion resistance is one of the main disadvantages
in magnesium and its alloys. However, the corrosion
resistance in high purity magnesium is not so poor but
comparable to those on die cast aluminum alloys or carbon
steels.1) Heavy metal impurities deterioratively affect corrosion resistance on magnesium.2) Therefore, in relatively new
magnesium alloys such as AZ91E alloy, heavy metal
concentrations are controlled to be low by metallographic
techniques in order to improve the corrosion resistance.1)
Limitation in heavy metal concentrations, however, decreases recycleablity in magnesium alloys,3) notwithstanding that
high recycleability is one of the main advantages in
magnesium alloy.
Thus, the authors have proposed a new technique for
improving the corrosion resistance without detriment to
recycleability, which is a deposition technique.4–7) In this
technique, magnesium or magnesium alloys are used for an
evaporation source and also for substrates. Since a purification process based on the retort effect is included in the
technique, the substrate is coated with deposited high purity
Table 1
magnesium layer. Advantages of the technique are as
follows: It is not necessary to remove the deposited layer
when the coated materials are recycled. Scraps with
magnesium alloys can be used for the evaporation source.
The process is not energy consuming one and non-toxic one
for environment.
In the present study, microstructures in coated layers
prepared by the optimized conditions for deposition, and also
in the layer fabricated on a substrate with large sizes are
investigated. Moreover, relationships between corrosion
resistance and microstructures are discussed based on in-situ
laser microscopic observation during salt immersion test.
2.
Experimental Procedures
Commercial grade pure magnesium and magnesium alloys
were used, chemical compositions of which are listed in
Table 1. Magnesium with 3N grade was used for an
evaporation source, while AZ31 magnesium alloy was used
for a substrate for deposition. Details of the deposition
technique were reported in the previous paper.6) In order to
obtain the optimum conditions for deposition, temperatures
Chemical compositions of magnesium and magnesium alloys (mass%).
Al
Zn
Mn
Si
Cu
Ni
Fe
3N–Mg
6N–Mg
0.004
0.000001
0.003
—
0.004
0.000005
0.005
0.000016
<0:001
<0:000005
<0:001
<0:000001
0.002
0.000002
AZ31
2.96
0.828
0.433
0.004
0.004
—
0.002
AZ91E
8.1
0.6
0.25
0.03
0.003
<0:001
0.003
*1Graduate
Student, Himeji Institute of Technology. Present address:
Yamato Scale Co., Ltd., Akashi, Japan.
*2Graduate Student, Himeji Institute of Technology. Present address:
Phoenix Electric Co., Ltd., Himeji, Japan.
High-purity Magnesium Coating on Magnesium Alloys by Vapor Deposition Technique for Improving Corrosion Resistance
3.
Results and Discussion
3.1 Deposition coating
Optimization of deposition conditions was carried out
using the previous furnace which is schematically illustrated
in Fig. 1(a). The optimum temperatures for the evaporation
900
(a)
Electric furnaces
Evac.
30
Stainless steel tube
Substrate
Evaporation
source
(b)
900
Temperature, T / K
of the substrate are changed in the range of 500 to 620 K with
constant temperature of about 980 K for the evaporation
source, which results in changing temperature profile in the
furnace. A new deposition coating furnace with large internal
sizes was developed based on the knowledge obtained in the
previous study, and coating a large substrate with about
60 100 6 mm3 in sizes was attempted.
Salt immersion tests were carried out to evaluate corrosion
resistance on the coated specimen together with un-coated
specimens for comparison in 3% NaCl solution at 300 K for
587 ks. Corrosion behavior on the specimens were investigated by in-situ observations using a laser microscope
(1LM15, Lasertec Co.) during immersion tests in 1% NaCl
solution at R. T. Specimens for in-situ observations were
about 10 10 5 mm3 in sizes. The specimens were
mechanically grounded, polished by cloth rubbing and
etched with an acetic picral etchant (acetic acid: 200 mL,
picric acid: 5 g, H2 O: 200 mL, ethanol: 100 mL). Surfaces of
the specimens were covered by an organic resin remaining a
surface for observation, and then dipped into the solution.
Details of the procedures were described in the previous
paper.8)
Microstructures of the corroded surface after immersion
tests were investigated by a scanning electron microscope
(SEM: S-900, Hitachi Ltd.) equipped with an energy
dispersive X-ray analyzer (EDX).
505
Position of
substrate
700
500
300
800
600
400
200
Distance form the Evaporation Source, x / mm
0
Fig. 1 Schematic illustration of the deposition furnace (a) and temperature
profile in the furnace (b).
source and the substrate were about 983 and 573 K,
respectively. The temperature profile is shown in Fig. 1(b).
There was a bump between the substrate and the evacuating
end. When a profile without the bump was applied for
deposition coating, the side surface opposite to the evaporation source was not coated.
Cross sections of the coated specimen fabricated with the
temperature conditions mentioned above under about
1:5 103 Pa for 7.2 ks are shown in Fig. 2. Thickness of
the deposited magnesium layer on the side surface opposite to
the evaporation source was about 6 mm (a) which was
comparable to those on the top surface, (b) and (c), and also
that on the side surface faced to the evaporation source (d).
Homogeneity in thickness has been improved compared with
the previous result,9) moreover defects such as pores or voids
(c)
(b)
Deposited
layer
Substrate
(d)
(a)
(e)
Deposited
layer
Substrate
Substrate
Mg
vapor
20 µm
Fig. 2 Cross sections of the coated specimen fabricated at 573 K for 7.2 ks. (a)–(d) show the cross sections taken at each position indicated
in (e).
506
H. Tsubakino et al.
Fig. 3
SEM images of the deposited surfaces (a)–(d) at the positions corresponding with (a)–(d) in Fig. 2, respectively.
at the interface between the substrate and the deposited layer
were diminished.
Microstructures of the deposited surfaces corresponding to
the cross sections in Figs. 2(a)–(d) are shown in Figs. 3(a)–
(d), respectively. Grain sizes varied from about 5 to 20 mm
depending on the surfaces, but any defects were not observed.
As reported in the previous paper,6) coating at higher
substrate temperature results in deposition of grains with
larger sizes. Small grains in the side surface opposite to the
evaporation source, (c), imply that the temperature was lower
or density of magnesium vapor was lower than those in other
surfaces. The latter case seems to be valid referring to the
temperature profile in Fig. 1.
The specimen sizes of the substrates used in the above
mentioned results were about 10 10 5 mm3 , which
seems to be too small for practical applications. Deposition
on large specimens with 20 40 6 and 10 40 6 mm3
in sizes were attempted under the same conditions as those
mentioned above. However, completely coated specimen
was not obtained, un-coated portions were remained. It was
considered to be due to the small internal sizes of the furnace,
about 30 mm in diameter and 900 mm in length. Therefore, a
large furnace has been newly developed, in which the internal
diameter was about 90 mm and the length was about
1000 mm. Only two furnaces were used for an evaporation
source and a substrate. Deposition on a substrate with about
60 100 6 mm3 in sizes was attempted under about the
same temperature profile to that in the small furnace
(Fig. 1(b)), however, it resulted in unsuccessful, the specimen
could not be coated. Coating of large area of the specimen
surfaces occurred at lower substrate temperatures than that
for the small furnace, about 400 K. It is considered that the
flow of magnesium vapor or density of magnesium vapor was
changed by increasing the internal diameter of the furnace.
However, the microstructures of the coated layer showed
columnar growth of grains, which are shown later. Since the
small internal size in the small furnace seems to play a
beneficial effect on coating, a condenser with an orifice of
about 10 mm in diameter was set between the evaporation
source and the substrate.
By using the orifice, deposition occurred at higher
temperatures compared with the cases without the orifice.
Cross section of the coated specimen prepared at 523 K for
7.2 ks without an orifice is shown in Fig. 4(a), the columnar
Fig. 4 Microstructures of the deposited layer fabricated using the large furnace at 523 K for 7.2 ks without an orifice (a) and at 573 K for
7.2 ks with an orifice (b).
High-purity Magnesium Coating on Magnesium Alloys by Vapor Deposition Technique for Improving Corrosion Resistance
507
(b)
(a)
50 µm
10 µm
Fig. 5 Surface microstructures of the same specimens in Figs. 4(a) and (b), that is, fabricated at 523 K for 7.2 ks without an orifice (a) and
at 573 K for 7.2 ks with an orifice (b).
growth of the grains can be seen. On the other hand, cross
section of the coated specimen prepared using an orifice at
the substrate temperature of 573 K for 7.2 ks, same as that in
the condition for the small furnace, the columnar growth
changed into planar growth as shown in Fig. 4(b). Microstructures of the surfaces on the same specimens as Figs. 4(a)
and (b) are shown in Figs. 5(a) and (b), respectively. It is
considered that there is a predominant crystal orientation for
columnar growth which appears at lower substrate temperatures. From a view point of improvement of corrosion
resistance, spaces between grains are not favorable.
(a)
Coated AZ31
3N-Mg
6N-Mg
10
AZ91E
Corrosion Rate, RC / mm year -1
20
AZ31
3.2 Corrosion behaviors
Corrosion rate on the coated specimen of the AZ31 alloy
fabricated at 573 K for 7.2 ks together with those on the uncoated 3N–Mg and 6N–Mg, and AZ31 and AZ91E alloys are
shown in Fig. 6, the values in which were calculated from
weight losses in salt immersion tests at 300 K for 587 ks using
a 3% NaCl solution. The corrosion resistance on the coated
specimen is superior to those on the un-coated 3N–Mg, AZ31
and AZ91E alloys, and comparable to that on the un-coated
6N–Mg.
In-situ observations of corrosion on the specimens can
interpret the reason why the corrosion rates on the un-coated
6N–Mg and the coated specimen are superior to those on
others. Figures 7(a) and (b) show the microstructures in the
un-coated 3N–Mg before and during a salt immersion test
using a 1% NaCl solution at R. T., respectively. There can be
seen many inclusions in the specimen which are indicated by
the arrows A, B and C in (a). When the specimen was
30
0
Fig. 6 Corrosion rates calculated from weight losses in a salt immersion
tests with 3% NaCl solution at 300 K for 587 ks. The specimens of 3N–Mg,
AZ31 alloy, AZ91E alloy and 6N–Mg were un-coated, while the coated
specimen was fabricated at 573 K for 7.2 ks in the small furnace.
immersed into the solution, bubbles were observed to be
formed at the position A, B and C in (b), that is, the
intermetallic inclusions act as the origins of evolution of
bubbles. These bubbles are believed to be hydrogen
bubbles10,11) evolved by the following corrosion reaction of
magnesium:12)
Mg ! Mg2þ þ 2e
ð1Þ
2H2 O þ 2e ! H2 þ 2(OH )
Mg
2þ
þ 2(OH ) ! Mg(OH)2
(b)
A B
C
A
B
C
500 µm
Fig. 7 Microstructure on the specimen of the un-coated 3N–Mg before immersion test (a), and in-situ observation of the same specimen
during immersion test (b). The arrows A-C in (a) and (b) indicate the same positions, respectively.
ð2Þ
ð3Þ
508
H. Tsubakino et al.
Rolling
direction
100 µm
Fig. 8 In-situ observation of corrosion reaction on the un-coated AZ31
alloy during a salt immersion test. Dark mist-like contrasts indicated by the
black arrows are due to flows of many fine bubbles. Rolling direction in
manufacturing process would be parallel to the white arrow.
In the case of the un-coated AZ31 alloy, bubble evolutions
were observed as shown in Fig. 8, in which mist-like
contrasts indicated by the arrows were due to flows of many
small bubbles. Although it is difficult to distingush between
bubbles and inclusions, alignment of positions for bubble
evolution (indicated by the arrows) along the rolling direction
(indicated by the white arrow) implies that inclusions provide
the origins for bubble formation.
On the bubble evolution in the un-coated AZ91E alloy, it
was observed to be formed at the second phase on a grain
boundary as shown in Fig. 9. Song et al.10) reported that the
bubbles were formed at the edge of (Mg17 Al12 ) phase,
which agree with the present result.
As reported in the previous papers,8,13) filiform corrosion
occurs on the un-coated 3N–Mg and the un-coated AZ31
alloy. In-situ observation of filiform corrosion on the uncoated AZ31 alloy is shown in Fig. 10(a). The filiform
corrosion occurred at the position indicated by the arrow and
then advanced remaining tails. Corrosion reaction occurred at
the heads of filiforms, while the tails were inactive. After the
in-situ observation, the specimen was rinsed and dried, then
observed by SEM. The area around the origin is shown in low
and high magnifications in (b) and (c), respectively. There
can be seen that the filiform corrosion occurred near the
inclusion indicated by the arrows. Energy dispersive X-ray
analysis showed that the inclusion was a compound with Al–
Mn system.
(a)
Figures 11(a) and (b) show about the same area on the uncoated AZ31 alloy after 1.86 and 1.92 ks from the start of
immersion, respectively. The head advanced about 180 mm
for 60 s, from the position indicated by the white arrow
toward the positions indicated by the arrows A and B.
Detailed observations showed that the filiform corrosion
proceeded along grain boundaries. There would be segregation of solute elements on grain boundaries. Inclusions and
grain boundaries are inhomogeneous microstructures in
metallographic meaning and also in electrochemical meaning, which provide predominant portions for corrosion
reactions.
The deposition method intrinsically includes purification
process, which results in coating the specimen surfaces with
high purity magnesium layer without intermetallic inclusions
or segregations. However, evolution of hydrogen bubbles
occurred as like as those in the un-coated 3N–Mg and the uncoated AZ31 alloy, although filiform corrosion did not occur
on the coated specimens as reported in the previous paper.8)
Figures 12(a) and (b) show the same area in the coated
surface in the salt solution. The bubble indicated by the arrow
in (b) was formed at the position indicated by the arrow in (a).
Careful observations showed that the same position on the
dried specimen (c), and then the position was found out by
SEM observation (d). High magnification image of the origin
on the bubble is shown in (e), fine flakes in which are
believed to be magnesium hydroxide formed by corrosion
reactions.14) Bubble evolution occurs at many positions at the
beginning of salt immersion test, the other part of the surface
is covered with magnesium hydroxide. The positions where
bubbles evolve are gradually covered by hydroxide with
proceeding immersion, and when the positions are completely covered, bubble formation is diminished as mentioned in
the previous paper.8) Magnesium and oxygen were detected
at the position of bubble formation as shown in Fig. 12(e).
Similar corrosion reactions to those on the coated specimen were observed on the un-coated 6N–Mg,8) that is, only
general corrosion occurred without filiform corrosion. Poor
corrosion resistance on the un-coated 3N–Mg and the uncoated AZ31 alloy are considered to be due to occurrence of
filiform corrosion. On the un-coated AZ91E alloy, filiform
corrosion did not occur, but internal regions surrounded by
grain boundary phase are observed to be predominantly
corroded, which agree with results reported by Song et al.10)
and Lunder et al.15)
(b)
100 µm
Fig. 9 In-situ observation of corrosion reaction in the un-coated AZ91E alloy. Hydrogen bubble indicated by the arrow in (b) were formed
at the vicinity of grain boundary indicated by the arrow in (a).
High-purity Magnesium Coating on Magnesium Alloys by Vapor Deposition Technique for Improving Corrosion Resistance
509
Fig. 10 In-situ observation of a filiform corrosion in the un-coated AZ31 alloy (a). SEM image with low and high magnifications (b) and
(c), respectively, of the area in (a) after salt immersion test. The arrows in (a), (b) and (c) indicate the origin of the filiform corrosion. EDX
spectrum taken at the inclusion indicated in (c).
(a)
(b)
A
A
B
B
100 µm
Fig. 11 In-situ observation of the filiform corrosion in the un-coated AZ31 alloy. After 1.86 and 1.92 ks from the start of the immersion,
(a) and (b), respectively. The filiform corrosion advanced from the position indicated by the white arrow in (a) to A and B through grain
boundaries.
Homogeneous microstructures with high purity and without second phases and segregations provide superior corrosion resistance for the coated specimens.
4.
Summary
Optimization of deposition conditions has lead to fabrication of coated layers on the substrate with homogeneous
thickness and microstructures. A large furnace has been
developed, which enables one to perform deposition on a
large substrate with 60 100 6 mm3 in sizes.
The corrosion resistance on the coated specimen is
superior to those on the un-coated 3N–Mg, un-coated AZ31
and AZ91E alloys, and comparable to that on the un-coated
6N–Mg.
In-situ observations on salt immersion tests using 1% NaCl
solution showed that intermetallic inclusions and grain
boundary segregations caused local corrosion such as filiform
corrosion on the un-coated 3N–Mg and the un-coated AZ31
alloy, and predominant grain interior corrosion on the uncoated AZ91E alloy. Homogeneous metallography on the
coated specimen leads to general corrosion, which results in
improving the corrosion resistance in magnesium alloys.
510
H. Tsubakino et al.
Fig. 12 In-situ observations of corrosion reaction in the coated specimen fabricated at 573 K for 7.2 ks in the small furnace; (a) and (b).
The bubble indicated by the arrow in (b) was evolved at the position indicated by the arrow in (a). Laser micrograph and SEM image
showing the position for bubble evolution, (c) and (d), respectively. High magnification SEM image and EDX spectrum taken at the
position for bubble evolution, (e) and (f), respectively.
Acknowledgements
This work was supported by the Priority Group of Platform
Science and Technology for Advanced Magnesium Alloys,
Ministry of Education, Culture, Sports, Science and Technology, Japan.
REFERENCES
1) K. N. Reichek and J. E. Hillis: SAE Technical Paper Series 850417
(1985) 1–12.
2) J. D. Hanawalt, C. E. Nelson and J. A. Peloubet: Trans. Metall. AIME.
147 (1942) 273–299.
3) H. Wentz and L. Ganim: Light Metal Age 50 (1992) No. 2, 14–21.
4) A. Yamamoto, A. Watanabe, K. Sugahara, T. Tsubakino and S.
Fukumoto: Scr. Mater. 44 (2001) 1039–1042.
5) H. Tsubakino, A. Watanabe, A. Yamamoto and S. Fukumoto: Mater.
Sci. Forum 350–351 (2000) 235–240.
6) A. Yamamoto, A. Watanbe, K. Sugahara, S. Fukumoto and H.
Tsubakino: Mater. Trans. 42 (2001) 1237–1242.
7) H. Tsubakino, A. Yamamoto, A. Watanabe, K. Sugahara and S.
Fukumoto: Proc. 4th Pacific RIM, (The Japan Inst. Metals, 2001)
pp. 1263–1266.
8) A. Yamamoto, A. Watanabe, K. Sugahara, S. Fukumoto and H.
Tsubakino: Mater. Trans. 42 (2001) 1243–1248.
9) H. Tsubakino, A. Yamamoto, A. Watanabe and S. Fukumoto: Materials
Week 2001 Proc. Inter. Cong. on Advanced Materials and Processings,
Munich, Germany, October 2001: CD-ROM, Ed. WerkstoffwochPartnerschaft GbR, Germany.
10) G. Song, A. Atrems, D. John and L. Zehn: Magnesium Alloys and their
Applications, (Wiley-VCH, Weinheim, 2000) 425–431.
11) J. D. Cotton: J. Electrochem. Soc. 136 (1989) 523C–527C.
12) G. L. Maker and J. Kruger: Inter. Mater. Rev. 38 (1993) 138–153.
13) A. Yamamoto, A. Watanabe, K. Sugahara, S. Fukumoto and H.
Tsubakino: Proc. 2nd Inter. Conf. on Environment Sensitive Cracking
and Corrosion Damage, ESCCD 2001, ed. by M. Matsumura et al.,
(Nichiki Printing, Hiroshima, 2001) pp. 160–167.
14) C. B. Baligaand and P. Tsakiropoulos: Mater. Sci. Technol. 9 (1993)
513–519.
15) O. Lunder, T. K. Aune and K. Nisancioglu: Corrosion 43 (1987) 291–
295.